Re: [Mops] AD review of draft-ietf-mops-treedn-03

["Eric Vyncke (evyncke)" <[e-mail]>]

Lenny, Chris, and Richard,

Just to be avoid a stalemate, were you expecting a reply of mine before submitting a revised I-D?

If so, then sorry about not replying: all the proposed changes were OK for me, i.e., you can proceed with a revised I-D without further wait.

Regards

-éric

From: Leonard Giuliano <[e-mail]>
Date: Wednesday, 3 April 2024 at 23:43
To: Eric Vyncke (evyncke) <[e-mail]>
Cc: MOPS Working Group <[e-mail]>, [e-mail] <[e-mail]>, [e-mail] <[e-mail]>
Subject: Re: AD review of draft-ietf-mops-treedn-03

Eric- thanks for the detailed review and suggestions.  See comments
inline:

On Tue, 2 Apr 2024, Eric Vyncke (evyncke) wrote:

|
| Dear authors, shepherd, and MOPS WG participants,
|
|
|
| Thanks for the work done on this document.
|
|
|
| As usual, please find below my AD review for this document. I am
| requesting either a revised I-D or a reply/explanation for all the
| points below before proceeding with the publication process (i.e., an
| IETF-wide Last Call).
|
|
|
|  *  Abstract: expand AR

Will expand

|
|  *  Abstract: please explain ?democratizing technology?
|

How about augmenting the last sentence of the abstract to:

"Finally, TreeDN is a decentralized architecture and a democratizing
technology in the way that it makes content distribution more accessible
to more people by dramatically reducing the costs of replication."

|
|
|  *  Section 1: a reference to "seven-second delay" would be welcome.
|

Will add ref to https://en.wikipedia.org/wiki/Broadcast_delay

|
|
|  *  Please remove section 1.1 as there is no normative language in this I-D.
|

Will remove

|
|
|  *  Section 3: s/ layer 3 hop/ layer-3 hop/
|

Will add

|
|
|  *  Section 3: ?firewall on the Internet ? unsure whether there are firewalls on the Internet (either delete ?firewall? or specify the location, e.g., ?at the edge?).
|

Taking an expansive view of the Internet here to include all L3 devices
between end hosts (sources and receivers).  Distinction between "edge" and
non-edge doesn't seem as germane here, plus a fw could theoretically sit
anywhere.  How about the following:

"To achieve ubiquitous availability on the global Internet, this
essentially means nearly every interface on every router and firewall
between all end hosts must support a multicast routing protocol like
PIM-SM [RFC7761] or mLDP [RFC6388]."

|
|  *  Section 4: suggest specifying that the global Internet is used as the underlay.
|

How about:

"That is, the replication service is available to users and applications
across the global Internet regardless of what protocols may exist in the
underlying networks that constitute the underlay."

|
|  *  Section 4: suggest using https://github.com/martinthomson/aasvg for a nicer SVG graphic.
|

Hmm, will work on that.

|
|
|  *  Section 4: expand AMT already in this section.
|

Can you elaborate?  AMT is first used in the first sentence of 4.1, and it
is expanded there.

|
|
|  * Section 4.1: ?users on unicast-only networks (AMT Gateways)? is ?(AMT
| gateways)? at the right location in this sentence or should it be
| phrased differently ? or s/users/residential gateways/ ? or both ?

AMT Gateways is correct. Probably better to s/ users/ end hosts/

|
|  *  Section 4.2 should there be a reference to figure 1 when mentioning ?on-net? ?
|

Well, Figure 1 is meant to provide an overall picture of the architecture
that is then detailed in 4.1 and 4.2.

|
|
|  *  Section 4.2: unsure whether a long digression on ASM vs. SSM fits this I-D. If so, then expand ASM.
|

Will expand ASM here.

|
|
|  *  Section 5 while I like the ?RaaS? concept, the actual term is not really defined in this section.
|

Well, this entire section (particularly the first para, and more
specifically the first sentence) is defining what RaaS is.

|
|
|  * Section 5: is there any operational drawbacks associated to TreeDN ?
| E.g., MTU reduction (due to tunnels), ... how to avoid ?leak? of paid
| content sent over mcast ? A forward reference to section 7 seems
|     required to me.

Can add the following to Sect 5:

"TreeDN introduces a deployment model that requires new considerations for
transport layer mechanisms that are frequently relied upon by traditional
unicast-based CDNs.  A discussion on these considerations and differences
can be found in section 7."


|
|  *  Section 7.2: CDN also tend to use HTTPS over QUIC (i.e., not over TCP).
|

Will re-word first sentence of 7.2 to:

"Traditional unicast-based CDNs frequently rely on HTTPS over TCP
transport and are...."

|
| Thanks again for the work done.
|
|
|
| -?ric
|
|
|
| PS: thanks also for your patience: I took some vacations before and after the IETF week...
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|